International Journal of Applied Research in Natural Products Vol. 7 (1), pp. 8-14. Directory of Open Access Journals ©2008-2014. IJARNP-HS Publication

Original Research Ultrasound-assisted extraction of volatile compounds from industrial Cannabis sativa L. inflorescences Da Porto C*, Decorti D, Natolino A Department of Food Science, University of Udine, via Sondrio 2/A, 33100 Udine, Italy Summary. This study investigated the use of ultrasound-assisted extraction (UAE) to recovery volatile compounds from the inflorescences of a fiber type Cannabis sativa L. cultivar. The results show that ultrasonic treatment not longer than 5 min allows to obtain an enhanced concentration of terpenes in comparison with maceration. Instead, an ultrasonic treatment longer than 5 min increased the concentration of δ-9-tetraidrocannabinol (THC). A preliminary screening of cannabis inflorescences scent was performed by headspace solid-phase microextraction (HS-SPME) with gas chromatography-mass spectrometry (GC-MS) avoiding the chemical modification and artifact formation that can occur in conventional methods . Industrial relevance. Inflorescences of fiber type Cannabis sativa cultivars are generally considered waste parts for fiber industry, although the inflorescences’ volatiles are pleasant to the human sensory system. Cannabis scent originate from volatile monoterpenes and sesquiterpenes . Traditionally, the recovery of floral fragrances from plants is by water distillation (hydro-distillation) or steam distillation to produce essential oils. However, these techniques take at least several hours and require the application of heating, which can produce the degradation of thermo labile compounds present in the starting plant material. Ultrasound-assisted extraction can be use as alternative method to extract aroma compounds from inflorescences of fiber type Cannabis sativa. The extracts so obtained could be used as ingredients for perfumes (cosmetic industry) or flavorings for beverages (food industry). Keywords. Ultrasound; Extraction; Cannabis sativa L.; terpenes; THC; HS-SPME

INTRODUCTION Cannabis sativa L. is an annual herbaceous plant belonging to the Cannabaceae family. The cultivation of this plant, native from Central Asia, and its use has been spread all over the world by man since thousands of years as a source of food, energy, fiber and medicinal or narcotic preparations (Wills, 1998; Russo, 2004; Jiang et al., 2006). Industrial hemp is a number of varieties of Cannabis sativa L. that are intended for agricultural and industrial purposes, differentiated from strains cultivated for medicinal use. They are cultivated for fiber and/or seed production. Only varieties of industrial hemp published by EU (Regulation (EC) No 1251/99 and subsequent amendments), are approved for planting in Europe. These varieties are eligible for cultivation only after the verification of their δ-9-tetrahydrocannabinol (THC) content, the principal psychoactive constituent of the cannabis plant, which must be less than 0.2% w/w (Regulation EC No. 1124/2008-12 November 2008). Today there is a renewed interest in cultivating industrial hemp for non-textile applications such as hemp-

Figure 1. Inflorescences of Cannabis sativa L.

___________________ *Corresponding Author.  [email protected]  +39 0432 558141 Available online http.//www.ijarnp.org

Da Porto et al.

based cosmetics and personal care products (i.e. soaps, shampoos, perfumes, etc) and health food including hempseed oil, flour and many food preparations in which they can be incorporated (i.e. bread, cookies, pancakes, etc.) (Mediavilla &. Steinemann, 2005; Leizer et al., 2000; Da Porto et al., 2012) Inflorescences of fiber type Cannabis sativa cultivars (Figure 1) are generally considered waste parts for fiber industry, although the inflorescences’ volatiles are pleasant to the human sensory system. Cannabis scent does not originate from the terpenophenolic cannabinoids, produced by glandular trichomes that occur on most aerial surfaces of the plant (Dayanandan & Kaufman, 1976; Turner et al., 1978), but from the more volatile monoterpenes and sesquiterpenes (Turner et al., 1980). Traditionally, the recovery of floral fragrances from plants is by water distillation (hydro-distillation) or steam distillation to produce essential oils. However, these techniques take at least several hours and require the application of heating, which can produce the degradation of thermo labile compounds present in the starting plant material. For the strain selection of drug type Cannabis sativa, many studies have been carried out on the essential oil (Mediavilla & Steinemann, 2005; Novak et al., 2001). Ultrasound-assisted extraction (UAE) has been proposed as alternative to conventional extraction, providing higher recovery of targeted compounds with lower solvent consumption and/or faster analysis and bioactivity properties. Its better extraction efficiency is related to the phenomenon called acoustic cavitation. When the ultrasound intensity is sufficient, the expansion cycle can create cavities or microbubbles in the liquid. Once formed, bubbles will absorb the energy from the sound waves and grow during the expansion cycles and recompress during the compression cycle. Further, bubbles may start another rarefaction cycle or collapse leading shock waves of extreme conditions of pressure and temperature (several hundred atmospheres and around 5000 K of temperature) (Figure 2) (Santos et al., 2009).

Figure 2 Creation of stable cavitation bubbles and creation and collapse of transient and stable cavitation bubbles. (a) Displacement (x) graph; (b) transient cavitation;(c) stable cavitation; (d) pressure (P) graph. (from Santos et al., 2009)

Thus, the implosion of cavitation bubbles can hit the surface of the solid matrix and disintegrate the cells causing the release of the desired compounds, so enhancing mass transfer and facilitating solvent access to the cell content (Vinatoru et al., 1999; Romdhane & Gourdan, 2002). This effect is much stronger at low frequencies (18–40 kHz). The benefit of using ultrasound in plant extraction has already been demonstrated for bioactive substances (Vinatoru et al., 1999) although few application are available concerning the extraction of aroma compounds (Caldeira et al., 2004; CabredoPinillos et al., 2006; Romdhane & Gourdan, 2002). To the best of our knowledge, there are no studies on the volatile compounds extracted by ultrasounds from the inflorescences of a fiber type Cannabis sativa cultivar and on the effect of this extraction method on the content of δ-9-tetrahydrocannabinol (THC). The aim of this study is to compare maceration and ultrasound-assisted extraction of volatile compounds from inflorescences of fiber type Cannabis sativa cultivar, generally considered waste parts for fiber industry, in order to evaluate the potential application of the extracts as ingredients for perfumes or flavorings for beverages. The extraction of δ9-tetrahydrocannabinol (THC) is also considered and discussed.

MATERIALS AND METHODS Plant Material. Fresh inflorescences of Cannabis sativa., cv. Felina (THC < 0.2%) were obtained from experimental trials carried out in Carnia (Friuli Venezia-Giulia region-Italy). On August 2012, from at least thirty plants of Cannabis sativa, the inflorescences were selected randomly from the cultivation area, handpicked and dried in the shade (moisture content 11.3% w/w ±1.1).

9

Volatiles from hemp inflorescences

Conventional maceration (M). An aliquot of 50 g of ground inflorescences were mixed with 250 mL of 70% ethanol v/v. The choice of ethanol-water solution (70% by volume) as extracting solvent was made based on its polarity relative to the aroma compounds of Cannabis sativa L. inflorescences and its acceptability for practical use. Maceration time was 3 h and stirring was performed during maceration. At the end of maceration time, the mixture was filtered under vacuum through Whatman No. 3, and the solvent was removed with a rotary vacuum evaporator. Each extraction was performed in triplicate using three different samples. Ultrasound-assisted extraction (UAE). Sonochemical experiments were carried out using an ultrasonic probe (Elettrofor Sonoplus model HD2200 with TT13FZ probe, Bandelin, Berlin; 20KHz working frequency; 200 W - amplitude setting displayed in % on the scale of 10 -100). The probe operated at 25 % of the scale according to the good results obtained in a previous work ( Da Porto et al., 2009). Three aliquots of 50 g of dried inflorescences, were added each with 250 mL of 70% ethanol v/v (extracting solvent). Each beaker and its content were immersed into 500 mL beaker containing ice-bath. The probe, submerged about 4 cm under the surface of the mixture, worked at 20 kHz frequency for 5 (UAE5), 10 (UAE10) and 15(UAE15) min. During extraction, the ice was replaced when melted to ensure a rapid dissipation of heating and temperature lower than 30°C. After being extracted, the mixtures were filtered under vacuum through Whatman No. 3 paper, and the solvent was removed with a rotary vacuum evaporator. Each extraction was performed in triplicate using three different samples. Static HS-SPME analysis coupled with GC-MS of inflorescences. An aliquot of 1 g of dried inflorescences was ground to powder and instantly introduced into a 10 mL vial. The vial was sealed by a Teflon septum and an aluminum cap. The SPME holder for manual sampling, and a 1 cm fibre assembly with 50/30µm divinyl/carboxen/polydimethylsiloxane coating (DVB/carboxen/PDMS) was purchased from Supelco Co. (Bellafonte, PA ). To study the equilibrium time profile of the inflorescences’ volatile compounds the preconditioned fibre (DVB/Carboxen/PDMS ) was inserted into the headspace of a set of vials maintained at 30°C for 5, 10, 15, 20, 30, 50 and 60 min. Thermal desorption of volatile analytes adsorbed on the SPME fibre was carried out in the GC injector port at 250 °C for 3 min. GC-MS analysis of the volatile compounds was performed using a Shimadzu gas chromatograph (model GC-17A) coupled to a Shimadzu mass spectrometer (model QP-5000). The fused silica column was a DB-5 fused-silica column (Supelco, Bellafonte, PA) (30 m x 0.25mm i.d., film thickness 0.25 µm). Working conditions were: injector 250 °C, transfer line to MS 250°C, oven temperature: start 45°C, hold 3 min; programmed from 45 to 190°C at 3°C min_1, hold 5 min, then further increase to 250°C at 20°C min_1, hold for 5 min ; carrier gas helium at flow rate 2.0 mL min -1 ; splitless; ionization: EI 70 eV; acquisition parameters: scanned m/z: 35–700 . Identification of the volatile compounds was carried out by comparing the Kovats retention indices determined by inserting a solution containing the homologous series of normal alkanes (C 7 -C 20 ) with those reported by literature (Bertoli et al., 2010; Fischedick et al., 2010) and with spectra of the NIST and WILEY libraries coupled with the software of GC-MS and Adams’ library (Adams, 2001). The repeatability of the GC-MS analysis was evaluated using the GC peak areas percent. GC-MS analysis of aromatic extracts. An aliquot of 5 mL of extract was added with 100 µL of n-dodecanol as internal standard and 1 µL injected in GC-MS. The volatile composition of the extracts was determined by direct GC-MS analysis. GC-MS analysis was performed under the same conditions reported for HS-SPME/GC-MS analysis of inflorescences (2.4) with the only exception of the injection mode which was with split ratio 1:40. Evaluation of δ-9-tetraidrocannabinol (THC ) content. The content of δ-9-tetraidrocannabinol (THC ) was determined following the method suggested by Cappelletto et al. (2001) Sample preparation. Inflorescences: an aliquot of 100 mg of ground inflorescences was added with 2 mL of absolute ethanol. The suspension was stirred at 50 °C over night. Subsequently the sample was centrifuged for 15 min at 3000 rpm at - 4 °C. The supernatant was separated and the solvent removed by nitrogen flow. The residue was dissolved with 200 µL of methanol with 0.5 mg mL -1 of squalene, as internal standard. On the dissolved residue THC determination was performed by GC-MS analysis. Extracts: from 0.5 mL of M, UAE5, UAE10 and UAE15 the solvent was removed by nitrogen flow. The samples thus obtained were analyzed under the same conditions reported for the inflorescences The analyzes were performed in triplicate. Direct GC-MS analysis. GC-MS analysis of δ-9-tetraidrocannabinol (THC) was performed using a Shimadzu gas chromatograph (model GC-17A) coupled to a Shimadzu mass spectrometer (model QP-5000). The fused silica column was a DB-5 fused-silica column (Supelco, Bellafonte, PA) (30 m x 0.25mm i.d., film thickness 0.25 µm). Working conditions were: injector 290 °C, transfer line to MS 300°C, oven temperature: start 230°C, hold 3 min; programmed from 230 to 280°C at 5°C min_1, hold 20 min.; carrier gas helium at flow rate 2.0 mL min -1; split ratio 1:15; ionization: EI 70 eV; acquisition parameters: scanned m/z: 35–700 . Identification was based on the matching of mass spectra of δ-9-tetraidrocannabinol with the reference mass spectra provided by the NIST library. Statistical analysis. Analysis of variance and Tukey’s test was conducted to identify differences among means for volatile compounds and THC. Statistical significance was declared at p≤0.05.

RESULTS AND DISCUSSION HS-SPME/GC-MS analysis of inflorescences. A preliminary screening for determination of the volatile compounds from inflorescences of Cannabis sativa L. using headspace solid-phase microextraction and gas chromatography–mass spectrometry was carried (Figure 3). SPME method was chosen to avoid the chemical modification and artifact formation that can occur in conventional methods. The main components identified from Cannabis sativa L. inflorescences, their retention indices and their percentage composition were summarized in Table 1. α-Pinene, myrcene, (E)-ocimene, terpinolene and caryophyllene were the most

10

Da Porto et al.

abundant volatiles released from flowers of Cannabis sativa L.. These findings were in agreement with those reported by Bertoli et al (2010) for the inflorescences of different fibre hemp varieties. The repeatability of the analysis was found to range between 2 and 9% RSD, indicating HS-SPME as an effective technique when applied to the analysis of cannabis flowers. However, unlike Bertoli et al. (2010) who performed HS-SPME analysis on cannabis inflorescences without reaching equilibrium, herein the equilibrium profile for the main volatile compounds of cannabis flowers was studied and determined (Figure 4).

Figure 3. HS-SPME/GC-MS analysis of volatile compounds from Cannabis sativa L. inflorescences. Peaks: 1.α- Pinene; 2. β-Pinene; 3. Myrcene; 4. (E)-ocimene; 5. Terpinolene; 6. Linalool; 7. Caryophyllene; 8. α-Humulene; 9. α-Cadinene

Table 1. Repeatability of the HS-SPME/GC-MS analysis of the volatile compounds of Cannabis sativa inflorescences

11

Volatiles from hemp inflorescences

This is an important distinction, since typically SPME is considered to be complete when the analyte concentration has reached distribution equilibrium between the sample matrix and the fiber coating (Zhang & Pawlisyn, 1993). The equilibrium was considered as achieved where the slopes of the curves – peak area (%) vs adsorption time- leveled off. As shown in Figure 4, the considered compounds achieved the equilibrium at 30°C after 50 minutes of adsorption. The more volatile compounds α-pinene and myrcene saturated the headspace more quickly than (E)-ocimene, terpinolene and caryophyllene, analytes with lower vapor pressures. 50 45 40

Peak area (%)

35 30

α-Pinene Myrcene

25

(E)-ocimene Terpinolene

20

Caryophyllene 15 10 5 0 0

10

20

30

40

50

60

70

Adsorption time (min )

Figure 4. Representative HS-SPME absorption time profiles of the major volatile compounds of Cannabis sativa L. inflorescences

Direct GC-MS analysis of aromatic extracts obtained by maceration and ultrasound assisted-extraction. The volatile composition of extracts obtained by maceration (M) and by ultrasound-assisted extraction for 5 (UAE5), 10 (UAE10) and 15 min (UAE15) are given in Table 2. By comparing Table 1 with Table 2, it can be observed that some volatile compounds detected by HS-SPME as α-pinene, α-phellandrene, 3-carene, α-terpinene, (E)-ocimene were not detected by direct GC-MS analysis of the extracts. This lost of volatile compounds could be due to their degradation as consequence of the extraction processes or/and to their poor solubilization in the ethanol-water solution. Table 2. Volatile composition of Cannabis sativa inflorescences extracts obtained by maceration (M) and by ultrasound-assisted extraction for 5 (UAE5), 10 (UAE10) and 15 min (UAE15)

Compound (mg mL )

Maceration M

β-Pinene Myrcene Limonene Terpinolene Linalool Caryophyllene α-Humulene α-Cadinene

0.20 d ± 3.32 0.37c ± 4.06 0.17c ± 9.28 0.12c ± 2.58 0.20d ± 8.03 2.08d ± 8.70 0.30b ± 9.26 0.54c ± 6.13

-1

a

Ultrasound-assisted extraction UAE5 UAE10 UAE15 0.55a ± 6.47 2.60a ± 6.75 0.80a ± 3.31 0.89a ± 3.35 0.42a ± 0.78 3.86a ± 3.45 0.46a ± 7.76 0.93a ± 5.95

0.34b ± 3.31 1.08b ± 5.12 0.48b ± 3.54 0.35b ± 3.47 0.31b ± 6.21 3.10b ± 1.53 0.45a ± 3.97 0.76b ± 1.36

0.27c ± 6.99 0.95b ± 1.86 0.42b ± 6.01 0.40b ± 8.63 0.28c ± 1.54 2.78c ± 0.18 0.42a ± 1.14 0.69b ± 0.11

Each data represents the mean of three extraction replicate ± RSD (%); a

Values with different letters within rows indicates significant differences(p < 0.05).

Compared with maceration, ultrasound-assisted extraction was found to enhance the recovery of monoterpene hydrocarbons, oxygenated terpenes and sesquiterpene hydrocarbons, as shown in Figure 5. When ultrasound was applied, extraction efficiency was generally improved due to the effect of ultrasonic cavitation which causes the intensification of mass transfer and thus closed interaction between the solvent and the inflorescences (Vinatoru & Toma, 1999). Beside the

12

Da Porto et al.

cavitational effects, the increase of recovery could be explained by the sound absorption property by ethanol-water mixture (Hemwimol et al., 2006). As illustrated in Figure 5, the highest concentrations of terpenes were obtained after 5 minutes of ultrasound application (UAE5). Sonication time longer than 5 minutes probably influenced the stability of these compounds, which also depends on their chemical structure.

6 a a

5 b b

mg /mL

4

c

3

Monoterpene hydrocarbons Oxygenated monoterpenes b

Sesquiterpene hydrocarbons

b

2

c

1

a

b

b

c

0 M

UAE5

UAE10

UAE15

Figure 5. Comparison of terpenes content from Cannabis sativa inflorescences extracted using maceration (M) and ultrasound-assisted extraction for 5 (UAE5), 10 (UAE10) and 15 min (UAE15) 1,6 a 1,4

ug /100 g dry matter

1,2

1

Tetrahydrocannabinol (THC)

0,8

0,6

0,4

b c

0,2

c

b

0 Inflorescences

M

UAE5

UAE10

UAE15

Figure 6.Comparison of δ-9-tetraidrocannabinol (THC) content between inflorescences and inflorescences extracts from Cannabis sativa L. obtained by maceration (M) and by ultrasound-assisted extraction for 5 (UAE5), 10 (UAE10) and 15 min (UAE15)

13

Volatiles from hemp inflorescences

δ-9-tetraidrocannabinol (THC) content of inflorescences and extracts. Among the cannabinoids, particular attention was focused in the present study to the content of δ-9-tetrahydrocannabinol (THC). As shown in Figure 6, the highest concentration of THC was detected in the inflorescences (0.0014% d.m.), though its concentration was very low respect to the limit of 0.2 % recommended by the Regulation EC No. 1124/2008 (12 November 2008). The average THC concentration of extracts was about 6-fold lower than the inflorescences. Compared to M, UAE5 was not significantly different for THC concentration (0.19 µg 100 mg d.m.-1 ± 1.52 vs 0.22 µg 100 mg d.m.-1± 0.48), as well as UAE10 compared to UAE15 (0.26 µg 100 mg d.m.-1± 0.43 vs 0.32 µg 100 mg d.m.-1± 0.49). However, THC concentration increased significantly when sonication time was longer than 5 minutes. This could be due to the fact that THC extraction rate is slow because it is located in the inner region of the trichomes particles.

CONCLUSIONS HS-SPME analysis allowed to perform a preliminary screening of the main volatile compounds of cannabis inflorescences avoiding the chemical modification and artifact formation that can occur in conventional methods. UAE was found to be an interesting alternative to maceration for extracting volatile compounds from cannabis inflorescence, but ultrasonic treatment must be not longer than 5 minutes to obtain an enhanced recovery of terpenes. Instead, an ultrasonic treatment longer than 5 minutes increased the concentration of δ-9-tetraidrocannabinol (THC). It can be concluded that the extract of Cannabis sativa inflorescences obtained by UAE carried out for 5 min could be used as ingredient for perfumes or flavoring for beverages. This could provide a complete utilization of industrial hemp using stems for fibre industry and inflorescences extracts for cosmetic and/or food industry.

REFERENCES Adams RP. 2001 Quadrupole mass spectra of compounds listed in order of their retention time on DB-5. In: Identification of essential oils components by gas chromatography/quadrupole mass spectroscopy Carol Stream, IL: Allured Publishing Bertoli A, Tozzi S , Pistelli L, Angelini LG 2010. Fibre hemp inflorescences: From crop-residues to essential oil production. Ind Crop Prod. 32 : 329–337 Cappelletto P, Brizzi M, Mongardini F, Barberi B, Sannibale M, Nenci G, Poli M, Corsi G, Grassi G,Pasini P. 2001. Italy-grown hemp: yield, composition and cannabinoid content. Ind Crop Prod 13: 101-113 Cabredo-Pinillos S , Cedron-Fernandez T , Gonzalez-Briongos M, Puente-Pascual L , Saenz-Barrio C. 2006. Ultrasound-assisted extraction of volatile compoundsfrom wine samples: Optimisation of the method, Talanta 69 : 1123–1129 Caldeira I, Pereira R ,Clımaco MC, Belchior AP, Bruno de Sousa R. 2004 Improved method for extraction of aroma compounds in aged brandies and aqueous alcoholic wood extracts using ultrasound. Anal. Chim. Acta 513 :125–134 Da Porto C, Voinovich D, Decorti D, Natolino A. 2012., Response surface optimization of hemp seed (Cannabis sativa L.) oil yield and oxidation stability by supercritical carbon dioxide. J Sup Fluids 68: 45-51 Da Porto C, Decorti D, Kikic I. 2009. Flavour compounds of Lavandula angustifolia L to use in food manufacturing: comparison of three different extraction methods. Food Chem 112: 1072-1078 Dayanandan P, Kaufman B. 1976. Trichomes of Cannabis sativa L. (Cannabaceae). Am. J. Bot. 63 : 578-591. Fischedick JT, Hazekamp A, Erkelens T, Hae Choi Y, Verpoorte R. 2010 Metabolic fingerprinting of Cannabis sativa L., cannabinoids and terpenoids for chemotaxonomic and drug standardization purposes. Phytochem 71 : 2058-2073 Hemwimol S, Pravasant P, Shotipruk A. 2006. Ultrasound-assisted extraction of anthraquinones from roots of Morinda citifolia. Ultrason. Sonochem. 13: 543-548. Jiang HE, Li X, Zhao YX, Ferguson DK, Hueber F, Bera S, Wang YF, .Zhao LC , Liu CJ , Li CS. 2006. A new insight into Cannabis sativa (Cannabaceae) utilization from 2500-year-old Yanghai Tombs, Xinjiang, China. J Ethnopharmacol 108: 414–422 Leizer C, Ribnicky D, Poulev A, Dushenkov S, Raskin I. 2006. The composition of hemp seed oil and its potential as an important source of nutrition. J. Nutraceut. Funct. Med. Foods 2 : 36–53. Mediavilla V, Steinemann S. 2005. Essential oil of Cannabis saliva L. strains. J. Agric. Food Chem. 53 : 424–429 Novak J, Zitterl-Eglseer K, Deans SG, Franz C. 2001. Essential oils of different cultivars of Cannabis sativa L. and their antimicrobial activity. Flavour Frag. J. 16 : 259-262. Romdhane M, Gourdan C. 2002. Investigation in solid-liquid extraction: influence of ultrasound, Chem. Eng. J. 87 : 11-19. Russo E. 2004. History of cannabis as a medicine. In: Guy GW ,Whittle BA, Robson PJ, editors. The medicinal uses of cannabis and cannabinoids, London (UK): Pharmaceutical Press, p. 1–16 Santos H.M., Lodeiro C., Capelo-Martinez J.L.(2009) The Power of Ultrasound In Capelo-Martinez J.L Ed .Ultrasound in Chemistry: Analytical Applications. WILEY-VCH Verlag GmbH & Co. KGaA, Weinhei, p.1-16. Turner CE, Hemphill P, Mahlberg G. 1978. Quantitative determination of cannabinoids in individual glandular trichomes of Cannabis sativa L. (Cannabaceae), Am. J. Bot 65 : 1103-1106 Turner CE, Elsohly MA, Boeren EG 1980. Constituents of Cannabis sativa L. XVII. A review of the natural constituents. J. Nat. Prod 43: 169-234. Vinatoru M, Toma M, Mason TJ 1999. Ultrasonically assisted extraction of bioactive principles from plants and their constituents. In: T.J. Mason TJ, editor, Advances in Sonochemistry, JAI Press: p.209-248 Wills S. 1998. Cannabis use and abuse by man: an historical perspective. In: Brown DT, editor. Cannabis: the genus Cannabis, Amsterdam: Harwood Academic Publishers p. 1–27 Zhang Z, Pawlisyn J. 1993. Headspace solid-phase microextraction. Anal Chem, 65: 1843-1852

14